Fixed point and Bregman iterative methods for matrix rank minimization (Q543413): Difference between revisions

The authors propose a fixed point continuation algorithm and a Bregman iterative algorithm for solving the linearly constrained nuclear norm minimization problem. The convergence of the fixed point iterative algorithm is established. A Monte-Carlo approximate singular value decomposition procedure is incorporated into the fixed-point continuation algorithm to improve the speed and its ability to recover low-rank matrices. Some numerical results are presented to show the effectively of the proposed algorithm.